1 @node SPSS Viewer Format
2 @section SPSS Viewer Format
4 SPSS Viewer or @file{.spv} files, here called SPV files, are written
5 by SPSS 16 and later to represent the contents of its output editor.
6 This section documents the format. This description is detailed
7 enough to read SPV files, but it is probably not sufficient to
10 An an aside, SPSS 15 and earlier versions use a completely different
11 output format based on the Microsoft Compound Document Format. This
12 format is not documented.
14 An SPV file is a Zip archive that can be read with @command{zipinfo}
15 and @command{unzip} and similar programs. The final member in the Zip
16 archive is a file named @file{META-INF/MANIFEST.MF}. This structure
17 makes SPV files resemble Java ``JAR'' files, but whereas a JAR
18 manifest contains a sequence of colon-delimited key/value pairs, an
19 SPV manifest contains the string @samp{allowPivoting=true}, without a
22 The rest of the members in an SPV file's Zip archive fall into two
23 categories: structure and details. ``Structure'' member names begin
24 with @file{outputViewer@var{nnnnnnnnnn}}, where each @var{n} is a
25 decimal digit, and end with @file{.xml}, and often include the string
26 @file{_heading} in between. Each of these members represents some
27 kind of output item (a table, a heading, a block of text, etc.) or a
28 group of them. The member whose output goes at the beginning of the
29 document is numbered 0, the next member in the output is numbered 1,
32 Structure members contain XML. This XML is sometimes self-contained,
33 but it often references other members in the Zip archive named as
37 @item @file{@var{prefix}_table.xml} and @file{@var{prefix}_tableData.bin}
38 @itemx @file{@var{prefix}_lightTableData.bin}
39 The structure of a table plus its data. Older SPV files pair a
40 @file{@var{prefix}_table.xml} file that describes the table's
41 structure with a binary @file{@var{prefix}_tableData.bin} file that
42 gives its data. Newer SPV files (the majority of those in the corpus)
43 instead include a single @file{@var{prefix}_lightTableData.bin} file
44 that incorporates both into a single binary format.
46 @item @file{@var{prefix}_warning.xml} and @file{@var{prefix}_warningData.bin}
47 @itemx @file{@var{prefix}_lightWarningData.bin}
48 Same format used for tables, with a different name.
50 @item @file{@var{prefix}_notes.xml} and @file{@var{prefix}_notesData.bin}
51 @itemx @file{@var{prefix}_lightNotesData.bin}
52 Same format used for tables, with a different name.
54 @item @file{@var{prefix}_chartData.bin} and @file{@var{prefix}_chart.xml}
55 The structure of a chart plus its data. Charts do not have a
58 @item @var{prefix}_model.scf
59 @itemx @var{prefix}_pmml.scf
60 Not yet investigated. The corpus contains only one example of each.
62 @item @var{prefix}_stats.xml
63 Not yet investigated. The corpus contains few examples.
66 The @file{@var{prefix}} in the names of the detail members is
67 typically an 11-digit decimal number that increases for each item,
68 tending to skip values. Older SPV files use different naming
69 conventions. Structure member refer to detail members by name, and so
70 their exact names do not appear to matter as long as they are unique.
72 @node SPV Structure Member Format
73 @subsection Structure Member Format
75 Structure members XML files claim conformance with a collection of XML
76 Schemas. These schemas are distributed, under a nonfree license, with
77 SPSS binaries. Fortunately, the schemas are not necessary to
78 understand the structure members. To a degree, the schemas can even
79 be deceptive because they document elements and attributes that are
80 not in the corpus and lack documentation of elements and attributes
81 that are commonly found in the corpus.
83 Structure members use a different XML namespace for each schema, but
84 these namespaces are not entirely consistent: in some SPV files, for
85 example, the @code{viewer-tree} schema is associated with namespace
86 @indicateurl{http://xml.spss.com/spss/viewer-tree} and in other with
87 @indicateurl{http://xml.spss.com/spss/viewer/viewer-tree} (note the
88 additional @file{viewer/} directory. In any case, the schema URIs are
89 not resolvable to obtain the schemas themselves.
91 One may ignore all of the above in interpreting a structure member.
92 The actual XML has a simple and straightforward form that does not
93 require a reader to take schemas or namespaces into account.
97 Parent: Document root or @code{heading} @*
98 Contents: [@code{pageSetup}] @code{label} [@code{container} | @code{heading}]*
100 The root of a structure member is a @code{heading}, which represents a
101 section of output beginning with a title (the @code{label}) and
102 ordinarily followed by content containers or further nested
103 (sub)-sections of output.
105 The document root heading may also contain a @code{pageSetup} element.
107 The following attributes have been observed on both document root and
108 nested @code{heading} elements:
111 @item Optional attribute: @code{creator-version}
112 The version of the software that created this SPV file. A string of
113 the form @code{xxyyzzww} represents software version xx.yy.zz.ww,
114 e.g.@: @code{21000001} is version 21.0.0.1. Trailing pairs of zeros
115 are sometimes omitted, so that @code{21}, @code{210000}, and
116 @code{21000000} are all version 21.0.0.0 (and the corpus contains all
117 three of those forms).
120 The following attributes have been observed on document root
121 @code{heading} elements only:
124 @item Optional attribute: @code{creator}
125 The directory of the software that created this SPV file,
126 e.g. @file{C:\PROGRA~1\IBM\SPSS\STATIS~1\22} or
127 @file{/Applications/IBM/SPSS/Statistics/22/SPSSStatistics.app/Contents/Resources/Java/../../bin}.
129 @item Optional attribute: @code{creation-date-time}
130 The date and time at which the SPV file was written, in a
131 locale-specific format, e.g. @code{Friday, May 16, 2014 6:47:37 PM
132 PDT} or @code{lunedì 17 marzo 2014 3.15.48 CET} or even @code{Friday,
133 December 5, 2014 5:00:19 o'clock PM EST}.
135 @item Optional attribute: @code{lockReader}
136 Whether a reader should be allowed to edit the output. The possible
137 values are @code{true} and @code{false}, but the corpus only contains
140 @item Optional attribute: @code{schemaLocation}
141 This is actually an XML Namespace attribute. A reader may ignore it.
144 The following attributes have been observed only on nested
145 @code{heading} elements:
148 @item Required attribute: @code{commandName}
149 The locale-invariant name of the command that produced the output,
150 e.g.@: @code{Frequencies}, @code{T-Test}, @code{Non Par Corr}.
152 @item Optional attribute: @code{visibility}
153 To what degree the output represented by the element is visible. The
154 only observed value is @code{collapsed}.
156 @item Optional attribute: @code{locale}
157 The locale used for output, in Windows format, which is similar to the
158 format used in Unix with the underscore replaced by a hyphen, e.g.@:
159 @code{en-US}, @code{en-GB}, @code{el-GR}, @code{sr-Cryl-RS}.
161 @item Optional attribute: @code{olang}
162 The output language, e.g.@: @code{en}, @code{it}, @code{es},
163 @code{de}, @code{pt-BR}.
167 Parent: @code{heading} or @code{container} @*
170 Every @code{heading} and @code{container} holds a @code{label} as its
171 first child. The root @code{heading} in a structure member always
172 contains the string ``Output''. Otherwise, the text in @code{label}
173 describes what it labels, often by naming the statistical procedure
174 that was executed, e.g.@: ``Frequencies'' or ``T-Test''. Labels are
175 often very generic, especially within a @code{container}, e.g.@:
176 ``Title'' or ``Warnings'' or ``Notes''. Label text is localized
177 according to the output language, e.g. in Italian a frequency table
178 procedure is labeled ``Frequenze''.
180 The corpus contains one example of an empty label, one that contains
184 Parent: @code{heading} @*
185 Contents: @code{label} [@code{table} | @code{text}]
187 A @code{container} serves to label a @code{table} or a @code{text}
191 @item Required attribute: @code{visibility}
192 Either @code{visible} or @code{hidden}, this indicates whether the
193 container's content is displayed.
195 @item Optional attribute: @code{text-align}
196 Presumably indicates the alignment of text within the container. The
197 only observed value is @code{left}. Observed with nested @code{table}
198 and @code{text} elements.
200 @item Optional attribute: @code{width}
201 The width of the container in the form @code{@var{n}px}, e.g.@:
206 Parent: @code{container} @*
207 Contents: @code{html}
209 This @code{text} element is nested inside a @code{container}. There
210 is a different @code{text} element that is nested inside a
211 @code{pageParagraph}.
214 @item Required attribute: @code{type}
215 One of @code{title}, @code{log}, or @code{text}.
217 @item Optional attribute: @code{commandName}
218 As on the @code{heading} element. For output not specific to a
219 command, this is simply @code{log}. The corpus contains one example
220 of where @code{commandName} is present but set to the empty string.
222 @item Optional attribute: @code{creator-version}
223 As on the @code{heading} element.
227 Parent: @code{text} @*
230 The cdata contains an HTML document. In some cases, the document
231 starts with @code{<html>} and ends with @code{</html}; in others the
232 @code{html} element is implied. Generally the HTML includes a
233 @code{head} element with a CSS stylesheet. The HTML body often begins
234 with @code{<BR>}. The actual content ranges from trivial to simple:
235 just discarding the CSS and tags yields readable results.
238 @item Required attribute: @code{lang}
239 This always contains @code{en} in the corpus.
243 Parent: @code{container} @*
244 Contents: @code{tableStructure}
247 @item Required attribute: @code{commandName}
248 As on the @code{heading} element.
250 @item Required attribute: @code{type}
251 One of @code{table}, @code{note}, or @code{warning}.
253 @item Required attribute: @code{subType}
254 The locale-invariant name for the particular kind of output that this
255 table represents in the procedure. This can be the same as
256 @code{commandName} e.g.@: @code{Frequencies}, or different, e.g.@:
257 @code{Case Processing Summary}. Generic subtypes @code{Notes} and
258 @code{Warnings} are often used.
260 @item Required attribute: @code{tableId}
261 A number that uniquely identifies the table within the SPV file,
262 typically a large negative number such as @code{-4147135649387905023}.
264 @item Optional attribute: @code{creator-version}
265 As on the @code{heading} element. In the corpus, this is only present
266 for version 21 and up and always includes all 8 digits.
271 Contents: @code{dataPath}
274 Parent: @code{tableStructure}
277 Contains the name of the Zip member that holds the table details,
278 e.g.@: @code{0000000001437_lightTableData.bin}.
281 Parent: @code{heading} @*
282 Contents: @code{pageHeader} @code{pageFooter}
285 @item Required attribute: @code{initial-page-number}
288 @item Optional attribute: @code{chart-size}
289 Always @code{as-is} or a localization (!) of it (e.g.@: @code{dimensione
290 attuale}, @code{Wie vorgegeben}).
292 @item Optional attribute: @code{margin-left}
293 @itemx Optional attribute: @code{margin-right}
294 @itemx Optional attribute: @code{margin-top}
295 @itemx Optional attribute: @code{margin-bottom}
296 Margin sizes in the form @code{@var{size}in}, e.g.@: @code{0.25in}.
298 @item Optional attribute: @code{paper-height}
299 @itemx Optional attribute: @code{paper-width}
300 Paper sizes in the form @code{@var{size}in}, e.g.@: @code{8.5in} by
301 @code{11in} for letter paper or @code{8.267in} by @code{11.692in} for
304 @item Optional attribute: @code{reference-orientation}
307 @item Optional attribute: @code{space-after}
313 Parent: @code{pageSetup} @*
314 Contents: @code{pageParagraph}*
319 Parent: @code{pageHeader} or @code{pageFooter} @*
320 Contents: @code{text}
322 Text to go at the top or bottom of a page, respectively.
325 Parent: @code{pageParagraph} @*
328 This @code{text} element is nested inside a @code{pageParagraph}. There
329 is a different @code{text} element that is nested inside a
332 The element is either empty, or contains cdata that holds almost-XHTML
333 text: in the corpus, either an @code{html} or @code{p} element. It is
334 @emph{almost}-XHTML because the @code{html} element designates the
336 @code{http://xml.spss.com/spss/viewer/viewer-tree} instead of an XHTML
339 The cdata can contain substitution variables: @code{&[Page]} for the
340 page number and @code{&[PageTitle]} for the page title.
342 Typical contents (indented for clarity):
345 <html xmlns="http://xml.spss.com/spss/viewer/viewer-tree">
348 <p style="text-align:right; margin-top: 0">Page &[Page]</p>
354 @item Required attribute: @code{type}
359 @node SPV Light Detail Member Format
360 @subsection Light Detail Member Format
362 A ``light'' detail member @file{.bin} consists of a number of sections
363 concatenated together, terminated by a byte 01:
366 light-member := header title styles dimensions data 01
369 The first section is a 0x27-byte header:
372 header := 01 00 version 01 (00 | 01) byte*21 00 00 table-id byte*4
377 @code{header} includes @code{version}, a version number that affects
378 the interpretation of some of the other data in the member. We will
379 refer to ``version 1'' and ``version 3'' members later on. It also
380 @code{table-id} is a binary version of @code{tableId} attribute in the
381 structure member that refers to the detail member. For example, if
382 @code{tableId} is @code{-4154297861994971133}, then @code{table-id}
383 would be 0xdca00003. The meaning of the other variable parts of the
387 title := value 01? /* @r{localized title} */
388 value 01? 31 /* @r{subtype} */
389 value 01? 00? 58 /* @r{locale-invariant title} */
390 (31 value | 58) /* @r{caption} */
391 int[n] footnote*[n] /* @r{footnotes} */
392 footnote := value (31 value | 58) byte*4
402 (i0 | i-1) (00 | 01) 00 (00 | 01)
404 byte[decimal] byte[grouping]
405 int[n-ccs] string*[n-ccs] /* @r{custom currency} */
408 x2 := 00 00 00 01 00 00 00 00 00 00 00 00 00 02 00 00 00 00 /* @r{18 bytes} */
410 styles2 := i0 /* @r{version 1} */
411 styles2 := count(count(x5) count(x6)) /* @r{version 3} */
412 x5 := byte*33 int[n] int*n
413 x6 := 01 00 (03 | 04) 00 00 00
414 string[command] string[subcommand]
415 string[language] string[charset] string[locale]
416 (00 | 01) 00 (00 | 01) (00 | 01)
418 byte[decimal] byte[grouping]
420 (string[dataset] string[datafile] i0 int i0)?
421 int[n-ccs] string*[n-ccs]
422 2e (00 | 01) (i2000000 i0)?
425 In every example in the corpus, @code{x1} is 240. The meaning of the
426 bytes that follow it is unknown.
428 In every example in the corpus, @code{x2} is 18 and the bytes that
429 follow it are @code{00 00 00 01 00 00 00 00 00 00 00 00 00 02 00 00 00
430 00}. The meaning of these bytes is unknown.
432 In every example in the corpus for version 1, @code{x3} is 16 and the
433 bytes that follow it are @code{00 00 00 01 00 00 00 01 00 00 00 00 01
434 01 01 01}. In version 3, observed @code{x3} varies from 117 to 150,
435 and its bytes include a 1-byte count at offset 0x34. When the count
436 is nonzero, a text string of that length at offset 0x35 is the name of
437 a ``TableLook'', e.g. ``Default'' or ``Academic''.
439 Observed values of @code{x4} vary from 0 to 17. Out of 7060 examples
440 in the corpus, it is nonzero only 36 times.
442 @code{encoding} is a character encoding, usually a Windows code page
443 such as @code{en_US.windows-1252} or @code{it_IT.windows-1252}. The
444 encoding string is itself encoded in US-ASCII. The rest of the
445 character strings in the file use this encoding.
447 @code{decimal} is the decimal point character. The observed values
448 are @samp{.} and @samp{,}.
450 @code{grouping} is the grouping character. The observed values are
451 @samp{,}, @samp{.}, @samp{'}, @samp{ }, and zero (presumably
452 indicating that digits should not be grouped).
454 @code{n-ccs} is observed as either 0 or 5. When it is 5, the
455 following strings are CCA through CCE format strings. Most commonly
456 these are all @code{-,,,} but other strings occur.
459 font := byte[index] 31 string[typeface]
461 (10 | 20 | 40 | 50 | 70 | 80)[f1]
465 (i0 | i2 | i64173)[f3]
466 (i0 | i1 | i2 | i3)[f4]
467 string[fgcolor] string[bgcolor]
469 (v3: int[f5] int[f6] int[f7] int[f8])
472 Each @code{font}, in order, represents the font style for a different
473 element: title, caption, footnote, row labels, column labels, corner
474 labels, data, and layers.
476 @code{index} is the 1-based index of the @code{font}, i.e. 1 for the
477 first @code{font}, through 8 for the final @code{font}.
479 @code{typeface} is the string name of the font. In the corpus, this
480 is @code{SansSerif} in over 99% of instances and @code{Times New
483 @code{fgcolor} and @code{bgcolor} are the foreground color and
484 background color, respectively. In the corpus, these are always
485 @code{#000000} and @code{#ffffff}, respectively.
487 The meaning of the remaining data is unknown. It seems likely to
488 include font sizes, horizontal and vertical alignment, attributes such
489 as bold or italic, and margins. @code{f1} is @code{40} most of the
490 time. @code{f2} is @code{i1} most of the time for the title and
491 @code{i0} most of the time for other fonts.
493 The table below lists the values observed in the corpus. When a cell
494 contains a single value, then 99+% of the corpus contains that value.
495 When a cell contains a pair of values, then the first value is seen in
496 about two-third of the corpus and the second value in about the
497 remaining one-third. In fonts that include multiple pairs, values are
498 correlated, that is, for font 3, f5 = 24, f6 = 24, f7 = 2 appears
499 about two-thirds of the time, as does the combination of f4 = 0, f6 =
503 font f1 f2 f3 f4 f5 f6 f7 f8
505 1 40 1 0 0 8 10/11 1 8
506 2 40 0 2 1 8 10/11 1 1
507 3 40 0 2 1 24/11 24/ 8 2/3 4
508 4 40 0 2 3 8 10/11 1 1
509 5 40 0 0 1 8 10/11 1 4
510 6 40 0 2 1 8 10/11 1 4
511 7 40 0 64173 0/1 8 10/11 1 1
512 8 40 0 2 3 8 10/11 1 4
516 dimensions := int[n-dims] dimension*[n-dims]
517 dimension := value[name]
525 int[n-categories] category*[n-categories]
528 @code{name} is the name of the dimension, e.g. @code{Variables},
529 @code{Statistics}, or a variable name.
531 @code{d1} is usually 0 but many other values have been observed.
533 @code{d3} is 2 over 99% of the time.
535 @code{d5} is 0 over 99% of the time.
537 @code{d6} is either -1 or the 0-based index of the dimension, e.g.@: 0
538 for the first dimension, 1 for the second, and so on. The latter is
539 the case 98% of the time in the corpus.
542 category := value[name] (terminal | group)
543 terminal-category := 00 00 00 i2 int[index] i0
546 @code{name} is the name of the category (or group).
548 @code{category} can represent a terminal category. In that case,
549 @code{index} is a nonnegative integer less than @code{n-categories} in
550 the @code{dimension} in which the @code{category} is nested (directly
553 Alternatively, @code{category} can represent a @code{group} of nested
557 group := (00 | 01)[merge] 00 01 (i0 | i2)[data]
558 i-1 int[n-subcategories] category*[n-subcategories]
561 Ordinarily a group has some nested content, so that
562 @code{n-subcategories} is positive, but a few instances of groups with
563 @code{n-subcategories} 0 has been observed.
565 If @code{merge} is 00, the most common value, then the group is really
566 a distinct group that should be represented as such in the visual
567 representation and user interface. If @code{merge} is 01, however,
568 the categories in this group should be shown and treated as if they
569 were direct children of the group's parent group (or if it has no
570 parent group, then direct children of the dimension), and this group's
571 name is irrelevant and should not be displayed. (Merged groups can be
574 @code{data} appears to be i2 when all of the categories within a group
575 are terminal categories that directly represent data values for a
576 variable (e.g. in a frequency table or crosstabulation, a group of
577 values in a variable being tabulated) and i0 otherwise, but this might
581 data := int[layers] int[rows] int[columns] int*[n-dimensions]
582 int[n-data] datum*[n-data]
585 The values of @code{layers}, @code{rows}, and @code{columns} each
586 specifies the number of dimensions represented in layers or rows or
587 columns, respectively, and their values sum to the number of
590 The @code{n-dimensions} integers are a permutation of the 0-based
591 dimension numbers. The first @code{layers} of them specify each of
592 the dimensions represented by layers, the next @code{rows} of them
593 specify the dimensions represented by rows, and the final
594 @code{columns} of them specify the dimensions represented by columns.
595 When there is more than one dimension of a given kind, the inner
596 dimensions are given first.
599 datum := int64[index] 00? value /* @r{version 1} */
600 datum := int64[index] value /* @r{version 3} */
603 The format of a datum varies slightly from version 1 to version 3: in
604 version 1 it allows for an extra optional 00 byte.
606 A datum consists of an index and a value. Suppose there are @math{d}
607 dimensions and dimension @math{i} for @math{0 \le i < d} has
608 @math{n_i} categories. Consider the datum at coordinates @math{x_i}
609 for @math{0 \le i < d}; note that @math{0 \le x_i < n_i}. Then the
610 index is calculated by the following algorithm:
614 for each @math{i} from 0 to @math{d - 1}:
615 index = @math{n_i \times} index + @math{x_i}
618 For example, suppose there are 3 dimensions with 3, 4, and 5
619 categories, respectively. The datum at coordinates (1, 2, 3) has
620 index @math{5 \times (4 \times (3 \times 0 + 1) + 2) + 3 = 33}.
623 value := 00? 00? 00? 00? raw-value
625 01 value-mod int32[format] double[x]
626 | 02 value-mod int32[format] double[x]
627 string[varname] string[vallab] (01 | 02 | 03)
628 | 03 string[local] value-mod string[id] string[c] (00 | 01)[type]
629 | 04 value-mod int32[format] string[vallab] string[varname]
630 (01 | 02 | 03) string[s]
631 | 05 value-mod string[varname] string[varlabel] (01 | 02 | 03)
632 | value-mod string[format] int32[n-args] arg*[n-args]
635 | int32[x] i0 value*[x + 1] /* @r{x > 0} */
638 A @code{value} boils down to a number or a string. There are several
639 possibilities, which one can distinguish by the first nonzero byte in
644 The numeric value @code{x}, presented to the user formatted according
645 to @code{format}, which is in the format described for system files.
646 @xref{System File Output Formats}, for details. Most commonly
647 @code{format} has width 40 (the maximum).
649 An @code{x} with the maximum negative double @code{-DBL_MAX}
650 represents the system-missing value SYSMIS. (HIGHEST and LOWEST have
651 not been observed.) @xref{System File Format}, for more about these
655 Similar to @code{01}, with the additional information that @code{x} is
656 a value of variable @code{varname} and has value label @code{vallab}.
657 Both @code{varname} and @code{vallab} can be the empty string, the
658 latter very commonly.
660 The meaning of the final byte is unknown. Possibly it is connected to
661 whether the value or the label should be displayed.
664 A text string, in two forms: @code{c} is in English, and sometimes
665 abbreviated or obscure, and @code{local} is localized to the user's
666 locale. In an English-language locale, the two strings are often the
667 same, and in the cases where they differ, @code{local} is more
668 appropriate for a user interface, e.g.@: @code{c} of ``Not a PxP table
669 for MCN...'' versus @code{local} of ``Computed only for a PxP table,
670 where P must be greater than 1.''
672 @code{c} and @code{local} are always either both empty or both
675 @code{id} is a brief identifying string whose form seems to resemble a
676 programming language identifier, e.g.@: @code{cumulative_percent} or
677 @code{factor_14}. It is not unique.
679 @code{type} is 00 for text taken from user input, such as syntax
680 fragment, expressions, file names, data set names, and 01 for fixed
681 text strings such as names of procedures or statistics. In the former
682 case, @code{id} is always the empty string; in the latter case,
683 @code{id} is still sometimes empty.
686 The string value @code{s}, presented to the user formatted according
687 to @code{format}. The format for a string is not too interesting, and
688 clearly invalid formats like A16.39 or A255.127 or A134.1 abound in
689 the corpus, so readers should probably ignore the format entirely.
691 @code{s} is a value of variable @code{varname} and has value label
692 @code{vallab}. @code{varname} is never empty but @code{vallab} is
695 The meaning of the final byte is unknown.
698 Variable @code{varname}, which is rarely observed as empty in the
699 corpus, with variable label @code{varlabel}, which is often empty.
701 The meaning of the final byte is unknown.
705 (These bytes begin a @code{value-mod}.) A format string, analogous to
706 @code{printf}, followed by one or more arguments, each of which has
707 one or more values. The format string uses the following syntax:
714 Each of these expands to the character following @samp{\\}. This is
715 useful to escape characters that have special meaning in format
716 strings. These are effective inside and outside the @code{[@dots{}]}
717 syntax forms described below.
720 Expands to a new-line, inside or outside the @code{[@dots{}]} forms
724 Expands to a formatted version of argument @var{i}, which must have
725 only a single value. For example, @code{^1} would expand to the first
726 argument's @code{value}.
728 @item [:@var{a}:]@var{i}
729 Expands @var{a} for each of the @code{value}s in @var{i}. @var{a}
730 should contain one or more @code{^@var{j}} conversions, which are
731 drawn from the values for argument @var{i} in order. Some examples
736 All of the values for the first argument, concatenated.
739 Expands to the values for the first argument, each followed by
743 Expands to @code{@var{x} = @var{y}} where @var{x} is the second
744 argument's first value and @var{y} is its second value. (This would
745 be used only if the argument has two values. With additional values,
746 the second and third values would be directly concatenated, which
750 @item [@var{a}:@var{b}:]@var{i}
751 This extends the previous form so that the first values are expanded
752 using @var{a} and later values are expanded using @var{b}. For an
753 unknown reason, within @var{a} the @code{^@var{j}} conversions are
754 instead written as @code{%@var{j}}. Some examples from the corpus:
758 Expands to all of the values for the first argument, separated by
761 @item [%1 = %2:, ^1 = ^2:]1
762 Given appropriate values for the first argument, expands to @code{X =
766 Given appropriate values, expands to @code{1, 2, 3}.
773 31 i0 (i0 | i1 string[subscript]) value-mod-i0-v1 /* @r{version 1} */
774 | 31 i0 (i0 | i1 string[subscript]) value-mod-i0-v3 /* @r{version 3} */
775 | 31 i1 int32[footnote-number] format
776 | 31 i2 (00 | 01 | 02) 00 (i1 | i2 | i3) format
777 | 31 i3 00 00 01 00 i2 format
779 value-mod-i0-v1 := 00 (i1 | i2) 00 00 int32 00 00
780 value-mod-i0-v3 := count(format-string
783 | 31 i0 i0 i0 i0 01 00 (01 | 02 | 08)
786 style := 01? 00? 00? 00? 01 string[fgcolor] string[bgcolor] string[font] byte
787 format := 00 00 count(format-string (58 | 31 style) 58)
788 format-string := count((i0 (58 | 31 string))?)